The Holiday Map Wars: Baidu Maps Keeps Getting Stronger, So Why Is It Drifting Further From Consumer Spending?

This National Day, two maps will face a stress test.

At a State Council Information Office press conference held on September 24, 2026, the Ministry of Transport estimated that during the seven-day National Day holiday, cross-regional travel by the whole society will reach 2.13 billion person-trips, averaging about 300 million per day; the first day is expected to see 340 million person-trips, with road travel accounting for more than 90 percent, and highway traffic is expected to reach 71 million vehicles.

Faced with this scale of travel, users must solve far more than just which road is faster. Whether charging guns are free, whether scenic area parking is subject to flow limits, whether restaurants are open, and where else they can go after their original plans fall through all test a map's understanding of the real-time world.

Before the long holiday, Amap and Baidu gave different answers. Amap incorporated behaviors such as "specially making a trip," "multiple repeat visits," and "locals navigating to a store" into its Street Scan List, configuring subsidies and business tools around merchants; Baidu Maps V22 launched mobile SR navigation, upgraded lane-level navigation and all-domain escort, and defined itself as a "spatiotemporal intelligence foundation" serving terminals such as cars, phones, and robots.

Baidu is also answering "where to go." V22 added AI travel, multi-destination itinerary linking, queries for scenic area opening hours and same-day foot traffic, and can also provide suggestions based on weather and real-time traffic conditions. Therefore, the divergence between the two companies is no longer about who can recommend destinations, but about how Baidu Maps and Amap make location data create value.

Amap has begun talking about store visits, repeat visits, and orders, while Baidu Maps focuses on showcasing roads, POIs, vehicle installations, and AI capabilities. What is truly worth asking is not whether Baidu can still do navigation well, but who is redefining the map business.

What the National Day test examines is no longer just navigation

During the 2025 National Day and Mid-Autumn Festival holiday, data monitored by the National Energy Administration showed that new energy vehicles on national highways charged 5.169 million times, with charging volume of about 122.87 million kWh, averaging 2.59 times the usual level for that year. For new energy vehicle owners, a map showing "there is a charger nearby" is only the passing grade; how long the queue is and whether a detour is worthwhile are the questions that truly affect the trip.

The destination itself may also temporarily become unavailable. Starting October 1, 2025, Nanjing Zhongshan Scenic Area implemented limited reservation parking for motor vehicles; during the National Day and Mid-Autumn Festival holiday that same year, the Jing County public security traffic management department made clear that when traffic on certain sections of the Wannan Sichuan-Tibet Line reaches preset peaks, measures such as diversion, turning vehicles back, and prohibition of passage may be taken.

Such scenarios require maps to handle dynamic capacity: where can people go, when can they go, and if they cannot get in, what alternatives remain. Whoever can influence these decisions stands before consumption occurs.

Amap's Street Scan List is trying to turn navigation behavior into evaluation signals. Whether users make a special detour to come, whether they visit the store again, and whether locals frequent it are used to judge merchant attractiveness and influence ranking on the list. The map thus begins to intervene in the distribution of offline traffic.

Baidu is also catching up on this lesson. According to disclosures by Baidu at the V22 launch, over the past year Baidu Maps' broad search requests grew 65 percent, and the adoption of AI proactive recommendations increased 4 times. However, the company did not further disclose how many store visits, repeat visits, or orders these recommendations generated.

This set of data proves that demand and feature usage have changed, but it cannot answer how recommendations translate into merchant operating results.

When maps begin to participate in destination choice, what map platforms control is no longer only the route entry point, but also where offline traffic may flow. The next question is which set of books the two companies plan to count this capability in.

After being merged into intelligent driving, whose results is Baidu Maps writing for?

The direction of Baidu Maps today can be traced from one organizational adjustment.

Baidu's fiscal year 2022 20-F explicitly disclosed that in 2022 the company integrated Baidu Maps into the intelligent driving business group, with the aim of creating synergy between the map App and map solutions for the automotive and transportation industries. That statement continued to appear in subsequent annual 20-F filings.

Organizational affiliation cannot prove that Baidu abandoned the consumer side, but it can affect business priorities. V21, released in 2025, already summarized its core capabilities as lane navigation, parking space navigation, 4D navigation, AI navigation, and intelligent-driving-grade data, and proposed opening data capabilities to OEMs and Tier 1 suppliers.

V22 continues this route. According to Baidu's disclosures, Baidu Maps has accumulated more than 13 million kilometers of true lane-level road data, more than 300 million POIs, receives more than 600 million user interaction feedbacks each month, and has mass-production installation experience in more than 60 million vehicles.

These are all company figures; installed vehicles do not equal active in-car users or continuously paying customers. But these data clearly present the results Baidu wants the outside world to see: road data, vehicle-side coverage, and terminal service capabilities.

Financial report disclosures form another contrast. After reviewing Baidu's five 20-F filings for fiscal years 2021 to 2025, all the documents introduce Baidu Maps, but no product-level indicators such as the App's MAU, DAU, independent revenue, orders, or navigation count were found. Not being listed separately does not mean there is no revenue or internal KPI, but it makes it difficult for outside investors to observe its independent operating results.

Apollo Go, which also belongs to the intelligent driving system, continues to disclose rides. Baidu's 20-F shows that it provided more than 1.5 million ride services in 2022; as of February 2026, cumulative public rides exceeded 20 million.

Apollo Go's rides certainly cannot be counted as Baidu Maps, but this disclosure difference is enough to show that the two businesses present different report cards in the capital narrative.

Four years after being merged into the intelligent driving business group, Baidu Maps' capabilities are increasingly counted into automobiles and AI terminals, while the user and commercial value created by the map App itself remains blurry in Baidu's public disclosures.

Amap has not won, but it is closer to merchants' money

Unlike Baidu, which merged Maps into the intelligent driving business group, Alibaba once placed Amap into the "To-Destination" framework of local life.

Alibaba's FY2024 annual report divided local life into "To-Home" and "To-Destination": the former is mainly handled by Ele.me, while the latter is mainly provided through Amap. Amap was defined as a one-stop entry point connecting navigation, local services, and ride-hailing. The same annual report said local life group revenue grew 19 percent year over year, mainly driven by order growth at Ele.me and Amap, but it did not break out the absolute volume and composition of Amap orders.

Amap's commercialization improvement subsequently entered Alibaba's financial reports. The FY2025 annual report showed that Amap achieved one profitable quarter in that fiscal year; the FY2026 annual report again confirmed Amap's revenue growth.

However, one profitable quarter does not equal stable profitability for the full year, and Alibaba also did not publish Amap's independent revenue, profit, GMV, and order scale.

After the 2025 organizational adjustment, Taobao Tmall, Ele.me, and Fliggy were integrated into the China e-commerce business group, while Amap was reclassified under "All others." This financial presentation cannot be directly regarded as a strategic downgrade, but it shows that Amap still maintains a relatively independent business affiliation.

What Amap has taken the initiative to do is not abandon the consumer-facing front end while doing enterprise technology services. In September 2026, Amap said the Street Scan List had covered 269 cities, served more than 880 million users cumulatively in its first year, and generated 36.6 billion kilometers of annual total navigation mileage to listed merchants; in 2025, orders at listed烟火 small shops grew by an average of 424 percent year over year.

These are all self-reported figures from Amap. Cumulative users cannot replace MAU, and the 424 percent figure also lacks public sample size, order definition, base period, and control group, so it cannot prove that growth was brought by the list. But this kind of disclosure already shows that Amap is trying to measure maps through merchant results.

Around the list, Amap also provides up to 4,800 yuan in annual merchant operating fee reductions, free aerial street views, AI store managers, and consumer subsidies. Amap is breaking a single arrival into metrics that can be managed: who came specially, who visited again, and whether traffic turned into orders.

The more Baidu looks like a foundation, the more it must explain who pays for this foundation

Baidu Maps is still continuously upgrading, and its navigation capabilities have not suddenly stopped working. The real dilemma occurs at another level.

V22 extends lane-level navigation to residential communities, industrial parks, and parking lots, mobile SR navigation provides spatialized road guidance, and all-domain escort supports more than 30 types of safety reminders. AI travel, hotel price comparison, and ride-hailing safeguards also show that Baidu has not exited the consumer side.

Choosing the "spatiotemporal intelligence foundation" also has industrial logic. Cars and robots both need road topology, dynamic events, and spatial understanding capabilities. If Baidu Maps becomes infrastructure relied on by more AI terminals, this market may not be smaller than mobile navigation.

The difficulty lies in how these capabilities translate into revenue. Five years of 20-F filings did not separately list map operating indicators; Baidu also did not explain how many of the more than 60 million mass-production installed vehicles generate continuous fees, or how many transactions AI travel, hotel, or ride-hailing services facilitated.

The radius of capabilities keeps expanding, yet independent operating results are difficult to observe.

Consumer ranking can directly affect merchant exposure, foot traffic, and budgets. When Amap uses navigation and repeat-visit behavior to judge whether a store is worth a special trip, what this mechanism competes for is already more than a list; it points to the power to distribute offline traffic.

Baidu has vast road, place, and travel data, yet it has not shown the outside world an equally clear merchant operating chain. Baidu Maps can more accurately understand where users want to go, but it has not yet proved how much observable commercial value this decision leaves behind.

The more Baidu Maps' industrial value grows, the less clearly its consumer-side commercial presence has become, and this is precisely the risk Baidu Maps faces.

Final thoughts

Baidu Maps' capabilities are increasingly counted into automobiles, intelligent driving, and AI terminals, while the independent results left on the consumer side are increasingly difficult for the outside world to see. This constitutes the real risk for Baidu Maps.

Amap is far from proving that it has won local life. At present there is only a single-quarter profit signal, independent revenue, profit, and order scale remain undisclosed, and the merchant growth disclosed by the Street Scan List also lacks a complete methodological explanation. But Amap has at least begun to answer something closer to money: why users set out, which store they arrive at, and whether they come back again.

Baidu is still proving how many devices maps can help understand the world, while Amap is trying to decide where offline traffic should go.

The two sides have not published transaction data under the same definition, so the location business cannot yet be said to have completed a change of hands. But when one map company keeps thickening its roads and spatiotemporal capabilities, and the other has already begun turning arrivals into merchant operating indicators, Baidu needs to further explain: for whom do these technical capabilities ultimately create revenue?

Baidu has not lost at navigation. What Baidu Maps may lose is the segment of value closest to consumption after users arrive.